Published in


Machine learning so clever it needs no training, now in UiPath

Photo by Ales Me on Unsplash

By Tommi Holmgren — Chief Product Officer at

RPA practitioners often face situations where they need to work with imperfect data. As a result, we see less than optimal automation rates or sometimes even abandoning a desired workflow as the estimated value creation does not justify the development cost.

We see these examples all the time. Here are some recent ones.

  • Document extraction was not able to get VAT category per line item reliably, so automation performance was low.
  • Purchase invoice processing automation was not able to complete workflows, as it did not know the cost center to assign invoices to. Humans still need to be involved all the time.
  • Sales lead assignment could not be automated in fear of losing valuable deals. Finding the most suitable sales rep for the lead requires non-deterministic judgement.

Do you have a similar example? We’d love to hear about it! We believe there are three things that would help users in many of cases like described above:

  1. to be able to predict a value, to fill the gaps of missing input data to the workflow
  2. to be able to predict a value that is necessary for the completion of automation in the most optimal way, but is not given in the input data
  3. to be able to achieve above things so fast you didn’t feel like machine learning was in the game.

Aito offers a ridiculously easy way to achieve all three and dramatically boost the automation rates. So far our customers have used our API or Python SDK to integrate Aito. If you haven’t heard about us before, start here.

We made a quick video how you can today use Aito in UiPath. Check it out! 👇

Using Aito machine learning in UiPath

What next, you might ask? We want to take this one step further, to completely annihilate the complexity of ML for RPA developers. And that means that we want to provide ready made UiPath activities for you to use. No JSON involed. No hassle with API calls involved.

What we are looking now is five UiPath users to test our ideas with. We want to ensure we get the experience right, and it really serves your needs. So if you have a use case where improving data will help, contact me at tommi[at] Once our activity is ready, these early users will get 50% discount for the first year of operation.

Originally published at




Imagine the future of data

Recommended from Medium

How I learnt how machines learn

Elon Musk — From the Innovation Breakthrough That Bridges Artificial Intelligence and Life Sciences

Network Whitepaper | Part 11 — Brain Computer Applications

Create VoiceOver for Videos using AI

The Tried and True Method for Hexapod Robot in Step by Step Detail

Why 2017 Isn’t Living Up to the AI Hype

3 Companies Using AI to Make the World Better

8 Movies Computer Scientists Will Enjoy

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store decision automation in the cloud. #ML for #nocode and #rpa operators.

More from Medium

Building an AI Traffic System With Python

How to Solve the Multi-Armed Bandit Problem: Epsilon-Greedy Algorithm

The GPL License Dependency Trap

Amazing, in AI training, one Nvidia 3090 can take on a large model with 18 billion parameters